Date:         Thu, 18 Mar 2004 14:53:16 -0500
Reply-To:     Sigurd Hermansen <HERMANS1@WESTAT.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Sigurd Hermansen <HERMANS1@WESTAT.COM>
Subject:      Re: Data
Comments: To: "julierog@ix.netcom.com" <julierog@ix.netcom.com>
Content-Type: text/plain

Roger:
I suspect that the longitude metaphor comes from the idea of observing
something once in each of a sequence of time zones (at different
longitudes). From prior work I have a habit of thinking of estimation errors
as longitudinal (perhaps distributed over long time lags) or variations due
to cross-section fixed effects or random effects. From that perspective, a
purely 'cross-section' study contrasts with a purely longitudinal study
(such as a birth cohort followed prospectively), although 'panel' studies
have both dimensions.
Sig

Subject:        Re: Data
Date:   Thu, 18 Mar 2004 13:28:02 -0500
From:   Roger Lustig <trovato@verizon.net>
Reply-To:       julierog@ix.netcom.com
To:     Jules Bosch <Jules@BOSCHSYSTEMS.COM>
Newsgroups:     comp.soft-sys.sas
References:     <001c01c40d0c$93901b00$6501a8c0@BSI>



Jules:
As latinate metaphors go, "longitudinal" is really pretty weak, longitude
having little to do with time per se, except that one needs a good
chronometer to determine one's longitude.  For "observed/measured/considered
over time", I prefer "diachronic," which comes with an equally pompous
antonym: "synchronic."

Roger

Jules Bosch wrote:
> Would someone please provide a brief description of longitudinal data?
> Is there an antonym, so to speak?
>
> TIA,
>
> Jules Bosch
> www.JBCDA.com